Don Greenberg developed a Kubeflow PyTorch MNIST pipeline example for the run-house/runhouse repository, focusing on end-to-end machine learning workflows. His work encompassed data preprocessing, GPU-accelerated model training, and automated inference deployment, leveraging Python and Shell scripting within a Kubeflow environment. To improve code maintainability and reduce technical debt, Don also removed outdated deployment examples for Mistral and Stable Diffusion XL on AWS Inferentia2, minimizing user confusion and risk of misconfiguration. The depth of his contribution is reflected in the integration of cloud computing best practices and deep learning techniques, emphasizing robust, production-ready ML pipeline development and codebase hygiene.